TB pixel clustering and cell clustering, 1 pixel clustering dataset (n20, 10 passes, rep1), 5 replicates of cell clustering (n20, 10 passes)

• First do pixel clustering, then do cell clustering using frequency of pixel clusters
• Gaussian blur sigma=2, k=20 (pixel metaclusters)
• For each pixel, divide each marker by total (each pixel sums to 1)
• k=20 (cell metaclusters)

Rep 1

Heatmap of pixel metacluster x cell cluster

Heatmap of marker x cell cluster (integrated)

Heatmap of marker x cell cluster (weighted)

Rep 2

Heatmap of pixel metacluster x cell cluster

Heatmap of marker x cell cluster (integrated)

Heatmap of marker x cell cluster (weighted)

Rep 3

Heatmap of pixel metacluster x cell cluster

Heatmap of marker x cell cluster (integrated)

Heatmap of marker x cell cluster (weighted)

Rep 4

Heatmap of pixel metacluster x cell cluster

Heatmap of marker x cell cluster (integrated)

Heatmap of marker x cell cluster (weighted)

Rep 5

Heatmap of pixel metacluster x cell cluster

Heatmap of marker x cell cluster (integrated)

Heatmap of marker x cell cluster (weighted)